Publications

We publish our research in high-impact conferences and journals within the field of Computer Science. We have collaborated top IT tech companies such as AWS AI Labs, Google Research, and NAVER AI Lab.
As of now, we published more than 30 CS top confernces in various domains, including NLP (EMNLP), CV (ICCV, CVPR), ML (NeurIPS, ICLR, ICML, AAAI), DM (KDD, CIKM, ICDM, WWW, SIGMOD).

Asterisk (*) denotes corresponding authors.

2023

  1. EMNLP
    Enhancing abstractiveness of summarization models through calibrated distillation
    Hwanjun Song, Igor Shalyminov, Hang Su, Siffi Singh, Kaisheng Yao, and Saab Mansour
    In Empirical Methods in Natural Language Processing, Findings, 2023
  2. EMNLP
    Fast and Robust Early-Exiting Framework for Autoregressive Language Models with Synchronized Parallel Decoding
    Sangmin Bae, Jongwoo Ko, Hwanjun Song*, and Se-Young Yun*
    In Empirical Methods in Natural Language Processing, Main, 2023
  3. NeurIPS
    Robust Data Pruning under Label Noise via Maximizing Re-labeling Accuracy
    Dongmin Park, Seola Choi, Doyoung Kim, Hwanjun Song, and Jae-Gil Lee*
    In Advances in Neural Information Processing Systems, 2023
  4. ICCV Oral
    Generating Instance-level Prompts for Rehearsal-free Continual Learning
    Dahuin Jung, Dongyoon Han, Jihwan Bang, and Hwanjun Song*
    In International Conference on Computer Vision, 2023
  5. ICML
    Context Consistency Regularization for Label Sparsity in Time Series
    Yooju Shin, Susik Yoon, Hwanjun Song, Dongmin Park, Byunghyun Kim, Jae-Gil Lee*, and Byung Suk Lee
    In International Conference on Machine Learning, 2023
  6. CVPR
    Re-thinking Federated Active Learning based on Inter-class Diversity
    SangMook Kim, Sangmin Bae, Hwanjun Song*, and Se-Young Yun*
    In International Conference on Computer Vision and Pattern Recognition, 2023
  7. ICLR
    Online Boundary-Free Continual Learning by Scheduled Data Prior
    Hyunseo Koh, Minhyuk Seo, Jihwan Bang, Hwanjun Song, Deokki Hong, Seulki Park, Jung-Woo Ha, and Jonghyun Choi*
    In International Conference on Learning Representation, 2023
  8. VLDB
    Data collection and quality challenges in deep learning: A data-centric ai perspective
    Steven Euijong Whang*, Yuji Roh, Hwanjun Song, and Jae-Gil Lee
    The VLDB Journal, 2023

2022

  1. NeurIPS
    Meta-Query-Net: Resolving Purity-Informativeness Dilemma in Open-set Active Learning
    Dongmin Park, Yooju Shin, Jihwan Bang, Youngjun Lee, Hwanjun Song*, and Jae-Gil Lee*
    In Advances in Neural Information Processing Systems, 2022
  2. NeurIPS
    Understanding cross-domain few-shot learning: An experimental study
    Jaehoon Oh, Sungnyun Kim, Namgyu Ho, Jin-Hwa Kim, Hwanjun Song*, and Se-Young Yun*
    In Advances in Neural Information Processing Systems, 2022
  3. ICDM
    Multi-view POI-level Cellular Trajectory Reconstruction for Digital Contact Tracing of Infectious Diseases
    Dongmin Park, Junhyeok Kang, Hwanjun Song, Susik Yoon, and Jae-Gil Lee*
    In International Conference on Data Mining, 2022
  4. CIKM
    FedRN: Exploiting k-Reliable Neighbors Towards Robust Federated Learning
    Sangmook Kim, Wonyoung Shin, Soohyuk Jang, Hwanjun Song*, and Se-Young Yun*
    In International Conference on Information and Knowledge Management, 2022
  5. CIKM
    e-clip: Large-scale vision-language representation learning in e-commerce
    Wonyoung Shin, Jonghun Park, Taekang Woo, Yongwoo Cho, Kwangjin Oh, and Hwanjun Song*
    In International Conference on Information and Knowledge Management, 2022
  6. CIKM
    ReFine: Re-randomization before Fine-tuning for Cross-domain Few-shot Learning
    Jaehoon Oh, Sungnyun Kim, Namgyu Ho, Jin-Hwa Kim, Hwanjun Song*, and Se-Young Yun*
    In International Conference on Information and Knowledge Management, 2022
  7. ICML
    Time Is MattEr: Temporal Self-supervision for Video Transformers
    Sukmin Yun, Jaehyung Kim, Dongyoon Han, Hwanjun Song, Jung-Woo Ha, and Jinwoo Shin*
    In International Conference on Machine Learning, 2022
  8. ICML
    Dataset condensation via efficient synthetic-data parameterization
    Jang-Hyun Kim, Jinuk Kim, Seong Joon Oh, Sangdoo Yun, Hwanjun Song, Joonhyun Jeong, Jung-Woo Ha, and Hyun Oh Song*
    In International Conference on Machine Learning, 2022
  9. CVPR
    Online continual learning on a contaminated data stream with blurry task boundaries
    Jihwan Bang, Hyunseo Koh, Seulki Park, Hwanjun Song, Jung-Woo Ha, and Jonghyun Choi*
    In International Conference on Computer Vision and Pattern Recognition, 2022
  10. TNNLS
    Learning from noisy labels with deep neural networks: A survey
    Hwanjun Song, Minseok Kim, Dongmin Park, Yooju Shin, and Jae-Gil Lee*
    IEEE Transactions on Neural Networks and Learning Systems, 2022
  11. AAAI
    Meta-learning for online update of recommender systems
    Minseok Kim, Hwanjun Song, Yooju Shin, Dongmin Park, Kijung Shin, and Jae-Gil Lee*
    In AAAI Conference on Artificial Intelligence, 2022
  12. AAAI Oral
    Covid-eenet: Predicting fine-grained impact of COVID-19 on local economies
    Doyoung Kim, Hyangsuk Min, Youngeun Nam, Hwanjun Song, Susik Yoon, Minseok Kim, and Jae-Gil Lee*
    In AAAI Conference on Artificial Intelligence, 2022

2021

  1. ICLR
    Vidt: An efficient and effective fully transformer-based object detector
    Hwanjun Song, Deqing Sun, Sanghyuk Chun, Varun Jampani, Dongyoon Han, Byeongho Heo, Wonjae Kim, and Ming-Hsuan Yang*
    In International Conference on Learning Representations, 2021
  2. ICLR
    Coherence-based label propagation over time series for accelerated active learning
    Yooju Shin, Susik Yoon, Sundong Kim, Hwanjun Song, Jae-Gil Lee*, and Byung Suk Lee
    In International Conference on Learning Representations, 2021
  3. BMVC
    Exploiting scene depth for object detection with multimodal transformers
    Hwanjun Song, Eunyoung Kim, Varun Jampan, Deqing Sun, Jae-Gil Lee, and Ming-Hsuan Yang*
    In British Machine Vision Conference, 2021
  4. NeurIPS
    Task-agnostic undesirable feature deactivation using out-of-distribution data
    Dongmin Park, Hwanjun Song, MinSeok Kim, and Jae-Gil Lee*
    In Advances in Neural Information Processing Systems, 2021
  5. KDD Oral
    Robust learning by self-transition for handling noisy labels
    Hwanjun Song, Minseok Kim, Dongmin Park, Yooju Shin, and Jae-Gil Lee*
    In International Conference on Knowledge Discovery and Data Mining, 2021
  6. KDD
    Machine learning robustness, fairness, and their convergence
    Jae-Gil Lee, Yuji Roh, Hwanjun Song, and Steven Euijong Whang*
    In International Conference on Knowledge Discovery and Data Mining, 2021
  7. AAAI
    Premere: Meta-reweighting via self-ensembling for point-of-interest recommendation
    Minseok Kim, Hwanjun Song, Doyoung Kim, Kijung Shin, and Jae-Gil Lee*
    In AAAI Conference on artificial intelligence, 2021

2020

  1. CIKM Oral
    Carpe Diem, Seize the Samples Uncertain" at the Moment" for Adaptive Batch Selection
    Hwanjun Song, Minseok Kim, Sundong Kim, and Jae-Gil Lee*
    In International Conference on Information and Knowledge Management, 2020
  2. ML
    Ada-boundary: accelerating DNN training via adaptive boundary batch selection
    Hwanjun Song, Sundong Kim, Minseok Kim, and Jae-Gil Lee*
    Machine Learning, 2020
  3. KDD Oral
    Hi-COVIDNet: Deep learning approach to predict inbound COVID-19 patients and case study in South Korea
    Minseok Kim, Junhyeok Kang, Doyoung Kim, Hwanjun Song, Hyangsuk Min, Youngeun Nam, Dongmin Park, and Jae-Gil Lee*
    In International Conference on Knowledge Discovery and Data Mining, 2020
  4. PAKDD
    Revisit Prediction by Deep Survival Analysis
    Sundong Kim, Hwanjun Song, Sejin Kim, Beomyoung Kim, and Jae-Gil Lee*
    In Pacific-Asia Conference on Knowledge Discovery and Data Mining, 2020
  5. WWW Oral
    TRAP: Two-level regularized autoencoder-based embedding for power-law distributed data
    Dongmin Park, Hwanjun Song, Minseok Kim, and Jae-Gil Lee*
    In The Web Conference, 2020

2019

  1. ICML Spotlight
    Selfie: Refurbishing unclean samples for robust deep learning
    Hwanjun Song, Minseok Kim, and Jae-Gil Lee*
    In International Conference on Machine Learning, 2019

2018

  1. SIGMOD Oral
    RP-DBSCAN: A superfast parallel DBSCAN algorithm based on random partitioning
    Hwanjun Song, and Jae-Gil Lee*
    In Special Interest Group on Management of Data, 2018

2017

  1. KDD
    PAMAE: parallel k-medoids clustering with high accuracy and efficiency
    Hwanjun Song, Jae-Gil Lee*, and Wook-Shin Han
    In International Conference on Knowledge Discovery and Data Mining, 2017